paul_craig law fact and discretion
TRANSCRIPT
-
8/10/2019 Paul_craig Law Fact and Discretion
1/34
Law, Fact and Discretion in the UK, EU and the USA
Professor Paul Craig
The elaboration and explication of the tests for judicial review of law, fact and discretion
in a particular legal system can, as all administrative lawyers will know, be a complex
task, since this is the very core of judicial review. Any attempt to explicate the tests for
judicial review of law, fact and discretion across three different legal systems is even
more difficult, but the enterprise can be rewarding for the comparative insights that it
thereby brings to the resolution of problems that endemic to all legal systems.
The scale of the task will however constrain the manner of explication. It is
clearly not possible within the confines of the space and time available this evening to
give a very detailed analysis of the salient issues. Nor would that be very helpful from the
perspective of comparative insight, since excessive detail would mask the similarities and
differences between the legal systems.
What follows will therefore necessarily be a schematic overview of the relevant
law in relation to review of law, fact and discretion, which is designed to reveal the
commonalities and differences between UK, EU and US law, and thereby set the stage
for discussion concerning the points of contact and contrast with French law.
1
-
8/10/2019 Paul_craig Law Fact and Discretion
2/34
1. JUDICIAL REVIEW: QUESTIONS OF LAW
1) Preliminary:
a) Legal systems will commonly distinguish between law, fact and discretion for the
purposes of judicial review. The nature of these distinguishing criteria can be
contentious. There are differing views, for example, as to the criteria for regarding
an issue as one of law, as opposed to fact. This is an endemic problem in all legal
systems, and the criteria to be found in any one legal system may not necessarily
be applied consistently by the courts. The criteria for distinguishing between law
and fact, or law and discretion may moreover differ as between legal systems. It
would clearly be impossible within the confines of the present discussion to
consider the detailed positive law for the three systems under scrutiny, since this
intellectual exercise could easily occupy the entirety of the time at our disposal.
b) For the purposes of the present discussion I shall therefore distinguish between
the terms law and fact in the following manner. The paradigm instance of fact
involves a dispute as to what are commonly termed primary facts, what people
saw, heard, did, etc., although as we shall see in the discussion concerning review
of fact, the nature of these inquiries may be more complex in certain areas than in
others. The paradigm instance of a question of law for the purposes of judicial
review in the UK, USA and EU is the meaning to be accorded to a statutory term
that defines the initial decision-makers scope of authority. Thus an agency may
be accorded power in the following terms. The empowering statute may say if an
employee is injured at work, then he shall be given compensation. The company
may contend that the person injured was not an employee and hence the agency
2
-
8/10/2019 Paul_craig Law Fact and Discretion
3/34
could not award compensation. The meaning of the term employee will
commonly be regarded as an issue of law in the flowing sense. The primary facts
will tell one that a person was, for example, working certain hours per week,
under certain terms, with a particular degree of control exercised by the employer.
These facts by themselves will not however tell one whether the person should
therefore be regarded as an employee for the purposes of the particular statute.
That can only be resolved through giving a legal meaning to the term employee.
The courts will interpret the term employee in the light of existing legal
principle and the purposes of the particular statute, which will lead to a legal
meaning for that term. The legal meaning will then be applied to facts, or to put
the matter conversely, it will then be decided whether the factual nature of the
working conditions of the person injured in this case meant that he came within
the definition of employee for the purposes of this statute.
2) The Test for Judicial Review: we can now consider the test for judicial review which
the courts will use in relation to issues of law. Four possible tests or types of test can
be distinguished in this regard.
a) The reviewing court simply substitutes judgment on the meaning of the relevant
legal term. By substitution of judgment I mean that the reviewing court will
decide for itself what the legal meaning of the term should be and will quash or
invalidate the decision of the initial decision-maker if it does not accord with the
meaning decided on by the court. This is the approach adopted in most, if not all
civil law countries; it is the approach adopted in the EU by the Community courts;
3
-
8/10/2019 Paul_craig Law Fact and Discretion
4/34
it is the predominant approach in the UK, (Anisminic, Page, South Yorkshire
Transport).
b) The reviewing court substitutes judgment on some of the issues of law that define
the scope of an agencys authority, but not all of them. This was the approach that
dominated the case law in the UK for 250 years under the title of the collateral or
jurisdictional fact doctrine.
c) The court substitutes judgment on certain issues of law, and uses a rationality test
for others. This is the approach in the USA in the famous Chevroncase.1In that
case the Supreme Court drew the following distinction. If a court reviewing an
agency's construction of the meaning of a term in a statute decided that Congress
had a specific intention on the precise question in issue then that intention should
be given effect to. The court will substitute judgment for that of the agency and
impose the meaning of the term Congress had intended.2 If, however, the
reviewing court decided that Congress had not directly addressed the point of
statutory construction, the court did notsimply impose its own construction on the
statute. Rather, if the statute was silent or ambiguous with respect to the specific
issue, the question for the court was whether the agency's answer was based on a
permissible construction of the statute. In answering this question the reviewing
court might uphold the agency finding even though it was not the interpretation
which the court itself would have adopted, and even though it was only one of a
range of permissible such findings that could be made.3Moreover, the Supreme
Court also held that the delegation to an agency of the determination of a
particular issue might well be implicit rather than explicit, and that in such
4
-
8/10/2019 Paul_craig Law Fact and Discretion
5/34
instances "a court may not substitute its own construction of a statutory provision
for a reasonable interpretation made by the administrator of an agency".4
d) The reviewing court uses a number of functional considerations to determine the
intensity of review of the issue of law placed before it in a particular case. This is
the approach used by the Canadian courts. They will take account of the expertise
of the agency, the nature of the question that is in dispute, whether that question is
one which the agency is well-placed to interpret, and whether there is a clause that
limits or restricts judicial review. The courts will, depending on the answers to
this range of factors, either substitute judgment, or more commonly use a less
intensive form of review, cast in terms of rationality.
3) Normative Reflection: The following points can be made by way of normative
reflection on this positive law.
a) The second, third and fourth approaches in the previous section are premised
implicitly or explicitly on the assumption that an issue can be characterized in
conceptual terms as one of law, without this necessarily entailing substitution of
judgment by the reviewing court. On these approaches, the reviewing court must
maintain some real control over the primary or initial decision-maker, but this
does not have to take the form of substitution of judgment, but can rather take the
form of a rationality test, or something akin thereto. This might well sound
shocking, heretical, or simply undesirable to those from a civil law background,
but then one of the virtues of comparative law is to test the fundamental
assumptions about ones own system against those from other legal orders.
5
-
8/10/2019 Paul_craig Law Fact and Discretion
6/34
b) This does not in any way mean that the first approach is wrong or misconceived.
There may be many reasons why a legal system feels that it should adhere to the
first approach. These reasons may well be eclectic: they may be in part derived
from a reading of that countrys constitution; they may in part flow from the view
that it is simply axiomatic that courts decide questions of law; they may be based
on the implicit legal theoretical assumption that it is wrong or dangerous to
acknowledge that legal terms may have more than one reasonable interpretation;
the established judiciary may be an unwilling to accept that an interpretation of a
legal term by the initial decision maker should ever be accorded legal respect;
there may be distrust of the established bureaucracy which translates for these
purposes into the desire to impose strong controls through substitution of
judgment on issues of law.
c) It should moreover be recognized that there are problems/tensions in legal
systems that use approaches two, three or four. This can be exemplified by the
jurisprudence of the US Supreme Court on the Chevron test. The court possesses
considerable judicial discretion in deciding whether a case should come within
part one of the test, and lead to substitution of judgment, or whether it should
come within part two of the test, and be subject to rationality review. There has
been real disagreement within the Supreme Court as to the nature of the divide
between the two parts of the test. The broad view of Chevron part one sees it as a
resolvability test: provided that the reviewing court can decide on the meaning of
the disputed term using ordinary tools of statutory construction broadly conceived
then the case fall within part one and the court substitutes judgment. The narrow
6
-
8/10/2019 Paul_craig Law Fact and Discretion
7/34
view of Chevron part one sees it as a clarity test: only if the meaning of the
disputed term really is clear from the face of the statute should the matter be
regarded as within Chevron part one. See Cardozo-Fonseca, Rust, Tobacco case.
See also the problems caused by theMeadjudgment.
2. JUDICIAL REVIEW: QUESTIONS OF FACT
1) Preliminary: Primary facts, what is seen done, heard etc, may be simple or they may
be complex. This may be obvious but it is important nonetheless. Where the primary
fact is relatively straightforward then it is realistic to think in terms of the primary
fact being established or not through observation or perception: did A hit B, was C
seen by D etc. However there may be many instances where more complex primary
facts require evaluation, rather than simple observation or perception. Thus
many of the factual issues that arise in, for example, competition cases, or those
concerning state aid, are of this kind.
2) The Standard of Review: There are once again a number of different tests for review
of fact that the courts can use.
a) The courts might substitute judgment on the facts for that of the primary decision-
maker. This option is acknowledged in UK and US law, and also implicitly in EU
law. It is however rare. The court will normally regard the primary decision-
maker, who may have heard oral argument and have intimate knowledge of the
dossier, as the person best equipped to make factual determinations. Thus in UK
and US law substitution of judgment will normally only be undertaken when there
is either some indication of structural infirmity with the decision-making process
7
-
8/10/2019 Paul_craig Law Fact and Discretion
8/34
used by the primary decision-maker, or where the reviewing court feels equally or
better equipped to find or evaluate facts as the primary decision-maker.
b) The courts can use a substantial evidence test. This is the general approach in US
law in relation to the assessment of more formal factual findings resulting in a
record of the proceedings. The courts will uphold the agency finding if there is
substantial evidence in the record as a whole, even if the court might not have
made those findings if it had been the initial decision-maker.5There must be such
relevant evidence as a reasonable mind might accept as adequate to support a
conclusion.
6
The courts might, alternatively, use a test for review couched in
terms of rationality or arbitrariness. Such tests have been deployed in the USA in
relation to factual findings resulting from less formal adjudicatory proceedings,
and in relation to informal rule-making, the argument being that criteria framed in
terms of rationality or arbitrariness are better suited to situations where there is no
formal record. There is however considerable support for the view that tests of
rationality and arbitrariness tend to converge with the substantial evidence test, in
the sense that a finding unsupported by substantial evidence is regarded as
arbitrary.7
c) The approach to review of factual findings in the UK has been uncertain and
unsatisfactory in the past. Some greater clarity has been forthcoming as a result of
a Court of Appeal judgment that laid down the following test: First, there must
have been a mistake as to an existing fact, including a mistake as to the
availability of evidence on a particular matter. Secondly, the fact or evidence must
have been established, in the sense that it was uncontentious and objectively
8
-
8/10/2019 Paul_craig Law Fact and Discretion
9/34
verifiable. Thirdly, the appellant (or his advisers) must not have been responsible
for the mistake. Fourthly, the mistake must have played a material (not
necessarily decisive) part in the Tribunals reasoning.
d) The test for review of fact in the EU is manifest error. It will be for the applicant
or claimant to show that the initial decision-maker, usually the Commission, made
some manifest error in its handling of the facts, and the Community courts state
repeatedly that it is not for them to substitute their judgment on the facts for that
of the Commission. However, as we shall below, the meaning accorded to
manifest error has altered considerably in the Courts more recent jurisprudence.
3) Normative Reflection: A number of points can be made about judicial approaches to
review of fact in the light of the tests set out in the preceding section.
a) Other things being equal, if a legal system opts for strict substitution of judgment
on issues of law, there will be greater incentive for the courts to manipulate the
law/fact or law /discretion distinctions for the following reason. The court may be
wedded to the idea that it should substitute judgment on issues of law, but feel
that it does not wish to do so in a particular case that has come before it. The court
does not feel that it can or should modify doctrinal orthodoxy in relation to
substitution of judgment on issues of law; it does not therefore regard it as open to
the court to admit that the issue is conceptually to be regarded as one of law, and
apply any form of review other than substitution of judgment. In such
circumstances, if the court does not wish to substitute judgment then its only
option is to classify the issue that has come before it as one of fact or discretion
9
-
8/10/2019 Paul_craig Law Fact and Discretion
10/34
and apply a less intensive standard of review. There is evidence of this in EU law
and in UK law.
b) It should be recognized that the courts can alter the intensity of review of facts,
even though they do not alter the doctrinal label through which such review is
conducted. This can be exemplified by the transformation in the meaning and
application review for manifest error in relation to facts in EU law. It should be
recognized at the outset that there has always been a duality or ambiguity latent in
the term manifest error: the controlling word manifest can either connote an
error that is both extreme and obvious, or it can connote an error that is serious,
even if not extreme, and even where the seriousness of the error only becomes
apparent after searching inquiry by the reviewing court. The story of review for
manifest error in relation to facts in EU law has been from the former meaning to
the latter, at least when reviewing for factual error in certain types of case, such as
those dealing with risk regulation and competition. Thus the approach in the
ECJs early jurisprudence was for very light touch review under the guise of
manifest error. If the alleged factual error was not apparent on the face of the
relevant decision or document, the ECJ would not intervene, and even if it was
apparent it would have to be very serious. In such cases, the ECJ would normally
devote no more than one or two paragraphs of its judgment to the issue. In the
cases on risk regulation or competition, review of facts is still undertaken through
a test of manifest error, but the court, especially the CFI, will undertake far more
searching scrutiny, often devoting thirty or more pages (not paragraphs) to the
issue, Pfizer, Airtours, Tetra Laval etc. There are indeed certain statements by the
10
-
8/10/2019 Paul_craig Law Fact and Discretion
11/34
ECJ in Tetra Laval that indicate that the reviewing court is to go beyond anything
that could readily be described as review for manifest error, and the reasoning
comes close to a form of substitution of judgment. Thus the ECJ stated that, not
only must the Community courts, inter alia, establish whether the evidence relied
on is factually accurate, reliable and consistent but also whether that evidence
contains all the information which must be taken into account in order to assess a
complex situation and whether it is capable of substantiating the conclusions
drawn from it.
c) It is important as a matter of principle to distinguish and keep separate two
related, but distinct, issues: the standard of proof required of the original decision-
maker for the finding of facts, and the standard of judicial review used by the
court in relation to those factual findings.
i) It is axiomatic that the existence or not of a factual error will be affected by
the standard of proof demanded in relation to the facts before the primary or
initial decision-maker makes the contested decision. There are a range of
standards from which to choose, including high degree of probability,
probability, possibility, sufficiency, a requirement that the evidence should be
convincing, or that there should be a preponderance of evidence to sustain the
action taken. It will normally be for the legislature to determine the standard
of proof in the enabling legislation, or if it does not do so the matter will be
decided by the courts.
ii) The second issue is, as stated above, the standard of judicial review used by
the reviewing court in deciding whether the standard of proof has been met or
11
-
8/10/2019 Paul_craig Law Fact and Discretion
12/34
not. The standard of proof tells us the degree of likelihood that must be
established in relation to factual findings in order for the primary decision-
maker to make its initial decision. It does not tell us the standard of judicial
review applied by the court in deciding whether the primary decision-maker
has met the standard of proof required of it. It is the latter that tells us how far
the reviewing court should reassess findings of fact made by the primary
decision-maker to decide whether the standard of proof for the initial decision
has been attained or not. The distinction between the standard of proof
required of the initial decision-maker and the standard of judicial review when
assessing whether the former has been met is important conceptually in legal
systems.8
iii) The standard of proof required of the primary decision-maker will frame the
test for review applied by the court, but the latter is nonetheless distinct from
the former. Thus it might be decided that the standard of proof for certain
administrative action should be probability, such that a chemical could only be
prohibited if it was probable that it would cause harm. This would still leave
open the standard to be applied by the reviewing court when determining
whether the facts and evidence before the administrative body sufficed to
establish the requisite probability. It might be felt that the test for review
should, for example, be cast in terms of substantial evidence: the court would
then consider whether the primary decision-maker had substantial evidence to
justify the conclusion that there was a probability of the chemical causing
harm such as to warrant Commission intervention.9
12
-
8/10/2019 Paul_craig Law Fact and Discretion
13/34
d) If the two issues, standard of proof required of the initial decision-maker and
standard of judicial review, are not kept distinct there is a danger that review of
fact will collapse into substitution of judgment. This may not be problematic if the
legal system has consciously decided that there should be substitution of
judgment for factual issues, but it is problematic where the legal system persists
with impression that review is based on manifest error. Consider once again the
preceding example, where the standard of proof demanded of the primary
decision-maker is that there should be a probability that a chemical might cause
harm before it could be banned. It is not for the reviewing court to decide
whether, if it had been the primary decision-maker, it would have concluded that
such a probability existed. This would be substitution of judgment by the
reviewing court for the view taken by the primary decision-maker. The reviewing
court must instead develop a standard of review that will allow it to assess
whether the person making the initial decision had enough evidence to warrant its
conclusion that the chemical would probably be harmful. The test for such review
might be that the evidence used by the decision-maker to justify the finding of
probability was, for example, substantial or sufficient, even if the reviewing court
might not itself have reached that conclusion had it been charged with the initial
decision. This is especially so given that the existence of facts and evidence might
be contestable, more particularly when the contested decision is a complex one.
13
-
8/10/2019 Paul_craig Law Fact and Discretion
14/34
3. JUDICIAL REVIEW: QUESTIONS OF DISCRETION
1) Preliminary: It is important at the outset to say something about the meaning of
discretion, which is a complex concept, on which there is a considerable
philosophical literature. Three senses of discretion can be distinguished for the
purposes of the present analysis, and all three can be found in the case law of the
courts in the UK, US, and EU.
a) The first type of case can for convenience be termed classic discretion. All grants
of power will take the form if X you may or shall do Y. It is common for there
to be multiple X conditions, so that the grant of power states that if X1, X
2, X
3, X
4
exists you may or shall do Y. Thus classic discretion is present where the relevant
statute, Treaty article, regulation directive or decision states that where certain
conditions exist the public body may take certain action. Put in terms of this
conceptual schema this is a case where the statute, Treaty article or regulation etc
states that if X1, X
2, X
3, X
4exists you may do Y.
b) The second type of case where the courts have held that discretion exists is where
there are broadly-framed conditions that have to be established before the power
or duty can be exercised at all. The phrase jurisdictional discretion can be used to
capture this type of situation. This is exemplified by Philip Morris Holland.10
The
issue before the ECJ concerned the meaning to be attributed to the phrases
abnormally low, and serious under employment within Article 87(3)(a) EC.
The ECJ held that the Commission had a discretion, the exercise of which
involved economic and social assessments that had to be made in a Community
context. Put in terms of the conceptual schema identified above, if X1, X
2, X
3, X
4
14
-
8/10/2019 Paul_craig Law Fact and Discretion
15/34
exists you may or shall do Y, the discretion recognized in Philip Morris Holland
resided in one of the X conditions. Article 87(3)(a) provides that if aid promotes
the economic development of areas where the standard of living is abnormally
low or where there is serious underemployment, the Commission may consider it
to be compatible with the common market. The dispute concerned the meaning to
be given to abnormally low, and serious under employment, these being
conditions that had to be met before the Commissions power to declare such aid
compatible with the common market became applicable. The ECJ was willing to
characterize these conditions as involving discretion, which then had an impact on
the standard of review applied.
c) There is also a third type of situation where discretion can exist, even where the
enabling statute, Treaty article, regulation, directive or decision is cast in
mandatory terms. There will of course be many instances where this is not so.
Where the relevant legislation provides if X, you shall do Y and Y entails a
specific measure there will be no room for any meaningful discretion once the
conditions have been met. There are however other instances where the content of
Y, the mandatory obligation, is cast in more general terms, thereby leaving some
measure of discretion as to how it should be fulfilled. This is exemplified by the
Common Agricultural Policy, Article 34 EC.
2) The Standard of Review in the UK:
a) The courts in the UK approach the review of discretion in two stages. They will
inquire, firstly, whether the discretion was exercised for improper purposes, and
whether it was based on relevant considerations. If the courts decide that the
15
-
8/10/2019 Paul_craig Law Fact and Discretion
16/34
discretion was exercised for improper purposes, or that it was based on irrelevant
considerations then the decision reached by the public body will be annulled. The
judicial exercise of deciding whether the power was exercised for improper
purposes etc is normally regarded as a matter of statutory interpretation, and the
courts will regard themselves as the arbiter of what the statute intended. They will
in that sense substitute judgment. The courts nonetheless also accept that the
determination of the meaning to be ascribed to complex or vague statutory words
may well be unclear or contentious.
b) The UK courts will, in addition, exercise control over the rationality of the
exercise of discretion. Thus even though the discretion may be deemed to have
been exercised for proper purposes and that it was based on relevant
considerations, it can still be challenged on grounds of rationality.
i) The basic conception of rationality review in the UK was set out in the
Wednesburycase:11
a decision could be attacked if it was so unreasonable that
no reasonable public body could have made it. To prove this would require
something quite extreme, Lord Greene MR giving the example of a teacher
being dismissed because of red hair. This distinctive meaning of the term
unreasonable was said to be warranted by the constitutional position of the
courts. They could not intervene simply because they believed that a different
way of exercising discretionary power would be more reasonable than that
chosen by the public body. This would be to substitute a judicial view for that
of the public body. Judicial intervention was therefore warranted only where
the decision could be shown to have been so unreasonable that no reasonable
16
-
8/10/2019 Paul_craig Law Fact and Discretion
17/34
public body could have made it. It is clear from Lord Greene MR's judgment
that he conceived of this test being used only in the extreme and hypothetical
instance of dismissal for red hair type of case. This was reinforced by Lord
Diplock in GCHQ who stated that this species of irrationality would only
apply to a decision which is so outrageous in its defiance of logic or of
accepted moral standards that no sensible person who had applied his mind to
the question could have arrived at it.12
ii) The courts have nonetheless applied the Wednesbury test to discretionary
decisions that could not, whether right or wrong, be classified as of the red
hair type.13
The test was applied in a way that made it closer to asking
whether the court believed that the exercise of discretion was reasonable. This
has become more explicit recognized in later cases:14
rationality review
covered not only decisions that defied comprehension, but also those made by
flawed logic.15
The loosening of Lord Greenes test received explicit
support from Lord Cooke in theITF case.16He regarded the formulation used
by Lord Greene as tautologous and exaggerated. It was not, said Lord Cooke,
necessary to have such an extreme formulation in order to ensure that the
courts remained within their proper bounds as required by the separation of
powers. He advocated a simpler and less extreme test: was the decision one
that a reasonable authority could have reached. Lord Cooke returned to the
topic in more forthright terms in Daly,17
where he said the following: [I]
think that the day will come when it will be more widely recognised that
Wednesbury was an unfortunately retrogressive decision in English
17
-
8/10/2019 Paul_craig Law Fact and Discretion
18/34
administrative law, insofar as it suggested that there are degrees of
unreasonableness and that only a very extreme degree can bring an
administrative decision within the legitimate scope of judicial invalidation.
The depth of judicial review and the deference due to administrative
discretion vary with the subject matter. It may well be, however, that the law
can never be satisfied in any administrative field by a finding that the decision
under review is not capricious or absurd.
iii) The UK courts have also varied the intensity with which they apply the
Wednesbury test in cases concerned with rights, prior to the Human Rights
Act 1998. The growing recognition of the importance of rights was
accommodated by modification of the meaning of unreasonableness. It is now
common to acknowledge that the courts apply the principles of judicial
review, including the Wednesbury test, with varying degrees of intensity
depending upon the nature of the subject-matter.18
Lord Bridge in Brind19
said that, in cases concerned with rights, the court must inquire whether a
reasonable Secretary of State could reasonably have made the primary
decision being challenged. The court should begin its inquiry from the
premise that only a compelling public interest would justify the invasion of
the right. Sir Thomas Bingham MRs formulation was very similar.20
The
court was to consider whether the decision was beyond the range of responses
open to a reasonable decision-maker, and the greater the interference with
human rights the more the court would require by way of justification.
18
-
8/10/2019 Paul_craig Law Fact and Discretion
19/34
c) The UK courts will also use proportionality as a test for review of discretion in
certain types of case. They will clearly do so in cases where EU law is concerned
and will also do so when reviewing discretionary determinations under the
Human Rights Act 1998. There is also now case law authority that proportionality
is the standard of review to be used when a public body seeks to resile from a
legitimate expectation. The courts have moreover also used proportionality, or
something very similar, in cases concerned with challenges to penalties.
Proportionality is not however thus far a free-standing, general principle of
administrative law in the UK.
3) The Standard of Review in the US: A similar development is evident in US law.
Under the Administrative Procedure Act 1946, agency findings can be set aside if
they are arbitrary, capricious or an abuse of discretion.21
a) The arbitrary and capricious test is therefore a principal tool for substantive
review of discretion. Judicial interpretation often matched the facial language of
the test. Plaintiffs faced an uphill task to convince a reviewing court that an
agency decision really was arbitrary and capricious, The criterion tended to be
narrowly interpreted, it being sufficient for the agency to show some minimal
connection between the statutory goal and the discretionary choice made by it;22
or to put the matter conversely the plaintiff would have to demonstrate some
manifest irrationality before the court would intervene. Thus as Shapiro states, in
fact in the 1940s and 50s, rules almost never failed the arbitrary and capricious
test,23
with New Deal judges being very reluctant to say that New Deal
19
-
8/10/2019 Paul_craig Law Fact and Discretion
20/34
bureaucrats had failed the APA sanity test, that is had done something arbitrary
and capricious.24
b) The label hard look developed because the courts began to desire more control
than allowed by this limited reading of the arbitrary and capricious test.25
In State
Farm26
the Supreme Court founded its intervention on the arbitrary and
capricious test, but then gave a broader reading to that phrase than that provided
in earlier cases. The court accepted that it should not substitute its judgment for
that of the agency. It could, however, intervene if any of the following defects
were present: if the agency relied on factors which Congress had not intended it to
consider; failed to consider an important aspect of the problem; offered an
explanation which ran counter to the evidence before the agency; was so
implausible that it could not be sustained; failed to provide a record which
substantiated its findings; or where the connection between the choice made by
the agency and the facts found was not rational. The hard look doctrine therefore
represented a shift from a previously more minimal substantive review, where
judicial intervention would occur only if there was serious irrationality, to one
where the courts would interfere where the broader list of defects set out above
are present. The hard look test proved to be a powerful tool, because of the
insistence on the provision of reasons, the demand for a more developed record
and a judicial willingness to assess the cogency of the reasoning process used by
the agency when it made its initial determination. This is not to say that the test
was unproblematic. There have been problems resulting from an excessive
20
-
8/10/2019 Paul_craig Law Fact and Discretion
21/34
demand for information and justification by the courts, which led some to coin the
phrase paralysis by analysis.
4) The Standard of Review in the EU: Similar judicial creativity is apparent in EU law.27
The general test under EU law is that factual and discretionary determinations will
only be set aside if there is some manifest error, misuse of power or a clear excess of
the bounds of discretion. This test has remained generally unchanged since the
inception of the European Community. The meaning given to the test has however
altered over time.
a) In the early years of the Communitys existence the test was applied with a light
touch. The ECJ repeatedly emphasized that its task was not to substitute
judgment for that of the Commission and that it would only intervene if there was
some patent error or misuse of power, more especially if the decision or rule
being challenged involved the complex balancing of factors that was normal in
the context of the Common Agricultural Policy.28
This test was applied with a
light touch, in the sense that the ECJ would commonly devote only one or two
paragraphs of its judgment to the issue, and would then conclude that no such
manifest error existed in the instant case.
b) The more modern case law reveals a rather different picture. The criteria for
review are formally the same. Manifest error, misuse of power or a clear excess of
the bounds of discretion remain the grounds of review, with the corollary that the
Community courts do not substitute judgment. It is however clear that while
retaining the established grounds of review the Community courts, and more
especially the Court of First Instance, the CFI, have been applying these with
21
-
8/10/2019 Paul_craig Law Fact and Discretion
22/34
greater intensity than hitherto, at least when reviewing certain types of action, in
particular risk regulation and competition cases. This is apparent from cases such
as Pfizer.29
One of the applicants arguments was that the Commission was in
breach of the test of manifest error in the way that it exercised its discretion. It
was therefore necessary for the CFI to decide whether the Community
institutions made a manifest error of assessment when they concluded that the
use of virginiamycin as a growth promoter constituted a risk to human health.30
The CFI proceeded in line with orthodoxy. It cited the well- established case law
that in matters concerning the common agricultural policy the Community
institutions had a broad discretion regarding the objectives to be pursued and the
means of doing so. Judicial review should therefore be confined to examining
whether the exercise of the discretion was vitiated by a manifest error, misuse of
power or clear excess in the bounds of discretion.31
The CFI also referred32
to
settled case law to the effect that where a Community authority was required to
make complex assessments in the performance of its duties, its discretion also
applied to some extent to the establishment of the factual basis of its action.33
It
followed said the CFI that in a case such as the present where the Community
institutions were required to undertake a scientific risk assessment and evaluate
complex scientific facts judicial review must be limited. The court should not
substitute its assessment of the facts for that of the Community institution, but
should confine its review once again to manifest error, misuse of power or clear
excess in the bounds of discretion.34
The CFI nonetheless considered in detail the
argument put forward by the applicant company. The CFI devoted 28 pages or 92
22
-
8/10/2019 Paul_craig Law Fact and Discretion
23/34
paragraphs of the judgment to this matter, which contrasts markedly with the 1 or
2 paragraphs to be found in the earlier jurisprudence. The same intensity of
review is apparent in recent cases concerned with judicial review of competition
decisions.35
It should be made clear that the CFI and ECJ have not always
reviewed for manifest error in the modern law with the intensity that is evident in
Pfizer and Airtours. There are many cases where they apply the test of manifest
error more intensively than in the early case law, but less intensively than in the
two cases considered above, with the consequence that there is in effect a
differential standard of review for discretion in EU law depending upon the nature
of the subject matter that is being reviewed by the courts.36
c) Proportionality is of course also a general ground for review of Community and
Member State action. It is commonly used to challenge discretionary
determinations made by the Commission, or by Member States when acting
within the confines of EC law. It is also well recognized that proportionality is a
test for review that can and is used with varying degrees of intensity by the
Community courts.37
Discretionary determinations may also be challenged on
other grounds, such as equality.
5) Normative Reflection: A number of points may be made by way of normative
reflection on the tests for review of discretion.
a) It is doubtful whether a legal system truly needs both rationality review and
proportionality review, for the following reasons.
i) First, if rationality review is limited to legal intervention only when the
decision being reviewed is so unreasonable that no administrative authority
23
-
8/10/2019 Paul_craig Law Fact and Discretion
24/34
would have reached it, then there is little doubt that this is a narrower form of
substantive review than that which normally exists under the three-part
proportionality test. If however the courts of a legal system really believe that
administrative action should only ever be annulled if it is irrational in the
extreme sense set out above, then it would be very strange for that system to
countenance proportionality review as an additional head of judicial review
operating alongside rationality review in this narrow sense. This is because
proportionality review would allow the courts to annul administrative action
in circumstances where it would not be open to challenge under the narrow
conception of irrationality. The logical solution would therefore be for the
courts of this legal system either to retain their narrow sense of rationality
review, refusing to review on grounds of proportionality, or to recognize and
apply proportionality as a general head of review, in which case the narrow
sense of rationality review would become redundant.
ii) Secondly, the same conclusion is true even if the courts of a legal system
apply rationality review in a more intensive manner. In such a system it would
also be unnecessary to have both rationality review and proportionality for the
following reason. The courts may well decide that while they should not
substitute their judgment for that of the administration, they should intervene
when the decision reached by the administration was not reasonable, not
simply where it was so unreasonable that no administrator could have reached
the decision. It would then be necessary for the courts to articulate in some
ordered manner the rationale for finding that an administrative choice was one
24
-
8/10/2019 Paul_craig Law Fact and Discretion
25/34
which could not reasonably have been made, where that choice fell short of
manifest absurdity. It is, however, difficult to see that the factors which would
be taken into account in this regard would be very different from those used in
the proportionality calculus. The courts would in some manner want to know
how necessary the measure was, and how suitable it was, for attaining the
desired end. These are the first two parts of the proportionality calculus. It is
also possible that a court might well, expressly or impliedly, look to see
whether the challenged measure imposed excessive burdens on the applicant,
the third part of the proportionality formula. If these kinds of factors are taken
into account, and some such factors will have to be, then it will be difficult to
persist with the idea that this is really separate from a proportionality test.
iii) Thirdly, there are considerable advantages to proportionality as a test for
review of discretion. It provides a structured form of inquiry. The three-part
proportionality inquiry focuses the attention of both the agency being
reviewed, and the court undertaking the review. The agency has to justify its
behaviour in the terms demanded by this inquiry. It has to explain why it
thought that the challenged action really was necessary and suitable to reach
the desired end, and why it felt that the action did not impose an excessive
burden on the applicant. If the reviewing court is minded to overturn the
agency choice it too will have to do so in a manner consonant with the
proportionality inquiry. It will be for the court to explain why it felt that the
action was not necessary etc in the circumstances. A corollary is that
25
-
8/10/2019 Paul_craig Law Fact and Discretion
26/34
proportionality facilitates a reasoned inquiry of a kind that is often lacking
under the traditional Wednesbury approach.
iv) Fourthly, the experience with proportionality in EU law reveals that the
concept is applied with varying degrees of intensity so as to accommodate the
different types of decision subject to judicial review. Cases involving rights
with be subject to the strictest review, cases involving broad discretion will
see the courts being more deferential to the administration and adopt lower
intensity proportionality review, while cases involving penalties with entail a
level of proportionality review intermediate between the other two.
38
b) We should also be aware of the inherent judicial creativity in the application of
the tests for review of discretion. It is the courts that will determine the real
meaning to be given to the criterion of review used by the particular legal system.
This is so whether review is characterized as being in terms of rationality,
arbitrary and capricious, manifest error or abuse of power.
i) The language of each of these terms, Wednesbury unreasonableness, arbitrary
and capricious and manifest error, would suggest that the court would only
intervene on these heads of review if the discretion were exercised in some
extreme manner. The reality has proven to be rather different in each of the
legal systems mentioned. The courts, while retaining the same head of review
in each legal system, have transformed them into a far more potent
mechanism for substantive review of administrative discretion. The precise
reasons why they have done this have varied in each legal system, but the
common denominator is that the courts in those systems wish to exert more
26
-
8/10/2019 Paul_craig Law Fact and Discretion
27/34
control over the exercise of discretion than hitherto; they wish to be able to
intervene and control discretionary determinations even if they are not
manifestly absurd or tainted by unreasonableness that is so extreme that no
reasonable authority would ever have reached that decision.
ii) It is axiomatic that if the courts operating in any system of administrative law
wish to expand the scope of substantive review of discretion they can do so in
one of two ways. They might choose to add new heads of substantive review
to those currently available within that system, a classic example being the
recognition and generalization of proportionality review. They might also
expand the reach of substantive review by taking existing heads of review and
giving them a more expansive interpretation than hitherto. Both techniques
might be used in tandem. The latter does however have attractions for the
judiciary. It is, other things being equal, easier for courts minded to expand
substantive review to preserve the impression of continuity with existing
doctrine if they continue to use well-recognized labels or heads of review,
while at the same time imbuing them with greater force than hitherto. This
approach obviates the need for the type of judicial self-inquiry that normally
attends the decision as to whether to introduce a new head of review to the
existing armoury. It will moreover often be the case that investing existing
heads of review with more vigour will only became apparent when the task
has been judicially accomplished. Reflection on the new status quo, whether
by academics or courts, will therefore take place against the backdrop of an
already developed jurisprudence that embodies the new or modified meaning
27
-
8/10/2019 Paul_craig Law Fact and Discretion
28/34
given to the classic head of review. It is in that sense ex post facto, as
compared with the judicial and scholarly discourse that will attend the
decision as to whether to introduce a new head of review, which will normally
be ex ante.
iii) It is equally important to understand the driving force behind this
development. If the courts within a legal system decide that they wish to
exercise more intensive rationality review, or that they desire to imbue an
arbitrary and capricious test with more force than hitherto, then a principal
tool used to achieve this end is to demand more reasons from the
administrative authority and to subject those reasons to closer scrutiny. They
may also demand more by way of inquiring into the evidentiary foundations
underlying the particular decision that is being challenged. The latter, the
evidentiary foundations for the contested decision, is related to the giving of
reasons, but is nonetheless separate. A legal duty to provide reasons imposed
by the courts will force the administrator to reveal why it acted in the way that
it did, and thereby renders it easier for the reviewing court to decide whether
those reasons were acceptable in the light of the relevant statute and its
objectives. The provision of reasons is therefore an essential tool when the
courts undertake substantive review of discretionary determinations, whether
through a test framed in terms of rationality, manifest error or arbitrary and
capricious. However a reviewing court may well feel that the reason that has
been given is indeed rational, provided that there is some evidence to
substantiate it. It is for this reason that reviewing courts will consider looking
28
-
8/10/2019 Paul_craig Law Fact and Discretion
29/34
to the evidentiary foundations of the contested decision. There is therefore a
proximate connection between procedural and substantive review. Thus
expansion of the duty to give reasons will normally lead to closer judicial
scrutiny of the administrations reasoning process in order to discover whether
there has been a substantive error. While it is perfectly possible in principle
for the courts to demand more by way of reasons, but still to engage in low
intensity review of the reasoning process in the context of substantive review,
the reality is that expansion of process rights will at the least encourage the
courts to engage in more intensive substantive review, because they have
more to work with and therefore feel more confident about asserting judicial
control.
iv) This still leaves open the issue touched on above, which is the reason or
reasons why the courts have sought to expand their control over discretionary
determinations. The reasons may of course differ from one legal system to
another, but there are nonetheless certain common themes. Thus, in the US,
the transformation of the arbitrary and capricious test from a relatively
minimal long stop to catch clear arbitrariness into a more potent tool for
substantive control over discretion, was motivated in part by an increasing
distrust of technical expertise combined with a greater willingness to engage
in more serious review of technocratic decision-making.39
This was combined
with increasing emphasis placed on the importance of transparency and
participation in the making of the initial decision or rule.40
This served to
place before a reviewing court a wider range of arguments about the content
29
-
8/10/2019 Paul_craig Law Fact and Discretion
30/34
of the contested norm, thereby facilitating closer review of the cogency of the
reasoning used by the agency.
v) We can see analogous considerations at work in the development of EU law.
The application of manifest error with a light touch in the early years was
likely influenced by the ECJs reticence in overturning norms in the new
Community order, more especially when they were brokered through hard
fought battles in the legislative arena. The EU is now firmly established and
the Community courts may well justifiably feel that it can withstand
annulment of some of its initiatives without thereby sending shock waves
through the system as a whole. More intensive deployment of manifest error
in relation to discretion and fact may also be explicable in terms of legitimacy.
There has as is well known been a growing discourse on the legitimacy of the
EU and on accountability of the decisions made therein. Imbuing the manifest
error test with greater force and thereby bolstering substantive review is one
way in which to enhance the accountability of those who made the initial
decision and hence to increase the legitimacy of the resulting norms. The
creation of the CFI is also undoubtedly of importance in this respect. It was
established to ease the burden on the ECJ. Its initial jurisdiction was for
complex cases with a heavy factual quotient, which required in-depth scrutiny
of facts and attention to the reasoning of the primary decision-maker. These
skills could be carried over when its jurisdiction was expanded to cover all
direct actions brought by non-privileged applicants. The CFI was therefore
well placed to put more intensive substantive review into practice. This is
30
-
8/10/2019 Paul_craig Law Fact and Discretion
31/34
reflected in the view expressed by Advocate General Vesterdorf that the
creation of the CFI as a court of both first and last instance for the
examination of facts in the cases brought before it was an invitation to
undertake an intensive review in order to ascertain whether the evidence on
which the Commission relies in adopting a contested decision is sound.41
Professor of English Law, St Johns College, Oxford.
1Chevron USA Inc vN.RD.C467 US 837 (1984).
2
See also, Etsi Pipeline Project v.Missouri 484 US 495, 517 (1988); Pittston Coal Group v. Sebben 488
US 105, 113 (1988); Bowen v. Georgetown University Hospital 488 US 204, 212 (1988); Adams Fruit
Company, Inc., v. Ramsford Barrett 494 US 638, 649 (1990);Department of the Treasury494 US 922
(1990);Dole v. United SteelworkersofAmerica 494 US 26, 43 (1990);NRDC v.Defense Nuclear Faclities
Safety Board 969 F.2d 1248, 1250-1251 (D.C.Cir.1992).
3467 US 837, 842-843.
4Ibid 844. See also,NLRB v.UnitedFood and Commercial Workers Union, Local 23, AFL-CIO 484 US
112, 123 (1987); Mississippi Power v.Moore 487 US 354, 380-382 (1988); Sullivan v.Everhart 494 US
83, 89 (1990); Pension Benefit Guaranty Corporation v. LTV Corporation 496 US 633, 648 (1990);
Department of the Treasury, Internal Revenue Service v. Federal Labour Relations Authority 494 US 922,
928 (1990);Armstrong World Industries, Inc., v.Commissioner of Internal Revenue 974 F.2d 422, 430 (3rd
Cir. 1992); West v. Sullivan 973 F.2d 179, 185 (3rd Cir.1992).
31
-
8/10/2019 Paul_craig Law Fact and Discretion
32/34
5A Aman and W Mayton,Administrative Law (West Publishing, 2nd. ed., 2001), 453-460.
6Consolidated Edison Co. v NLRB 305 U.S. 197, 229 (1938).
7
Associated Industries v US Dept of Labor 487 F.2d. 342, at 350 (2d Cir. 1973); Association of Data
Processing Service Organizations, Inc. v Board of Governors of the Federal Reserve System 745 F.2d 677,
683 (D.C. Cir. 1984).
8A. Aman and W. Mayton, Administrative Law (West, 2nded., 2001), 234-6; Steadman v. SEC 450 US 91
(1981).
9Ibid. at 453-71.
10Case 730/79, Philip Morris Holland BV v. Commission[1980] ECR 2671.
11Associated Picture Houses Ltd.v.Wednesbury Corporation[1948] 1 K.B. 223, 228-230.
12Council of Civil Service Unionsv.Minister for the Civil Service[1985] A.C. 374, 410.
13Hall & Co. Ltd. v. Shoreham-by-Sea Urban District Council [1964] 1 W.L.R. 240; R. v.Hillingdon
London Borough Council, ex p. Royco Homes Ltd.[1974] Q.B. 720; West Glamorgan County Council v.
Rafferty [1987] 1 W.L.R. 457; R. v. Secretary of State for the Home Department, ex p. Tawfick [2001]
A.C.D. 28;R. v. Secretary of State for Health, ex p. Wagstaff [2001] 1 W.L.R. 292;R. (on the application
of Howard) v. Secretary of State for Health [2002] E.W.H.C. 396; R. (on the application of Von
Brandenburg) v. East London and the City Mental Health Trust [2002] Q.B. 235. .
14R. v. Lord Saville of Newdigate, ex p. A [1999] 4 All E.R. 860, para. 33; R. v. Parliamentary
Commissioner for Administration, ex p. Balchin [1997] C.O.D. 146; R. v. North and East Devon Health
Authority, ex p. Coughlan[2001] Q.B. 213, para 65.
15Ibid. para. 65.
16R. v. Chief Constable of Sussex, ex p. International Trader's Ferry Ltd.[1999] 2 A.C. 418, 452.
17R. v. Secretary of State for the Home Department, ex p. Daly [2001] 2 A.C. 532, 549; R. (on the
application of Louis Farrakhan) v. Secretary of State for the Home Department [2002] 3 W.L.R. 481, para.
66.
18Sir John Laws, Wednesbury, in C. Forsyth and I. Hare (eds.), The Golden Metwand and the Crooked
Cord, Essays in Honour of Sir William Wade (Oxford University Press, 1998), pp. 185-202.
32
-
8/10/2019 Paul_craig Law Fact and Discretion
33/34
19R. v.Secretary of State for the Home Department, ex p. Brind [1991] 1 A.C. 696, 748-749.
20R. v.Ministry of Defence, ex p. Smith [1996] Q.B. 517.
21
Administrative Procedure Act 1946, s. 706(2)(a).
22A. Aman and W. Mayton, Administrative Law (West, 2nd ed., 2001), 519-29; S. Breyer, R. Stewart, C.
Sunstein and M. Spitzer, Administrative Law and Regulatory Policy (Aspen Law & Business, 5th ed.,
2002), 415.
23M. Shapiro, Codification of Administrative Law: The US and the Union (1996) 2ELJ 26, 28.
24Ibid. at 33.
25See, Greater Boston Television Corp.v.Federal Communications Commission444 F.2d 841, at 850-53
(D.C.Cir. 1970), cert. denied 403 U.S. 923 (1971); Environmental Defense FundInc v. Ruckelshaus439
F.2d 584 (D.C.Cir. 1971); H. Leventhal, Environmental Decision making and the Role of the Courts 122
U.Pa LRev509 (1974); R. Stewart, The Reformation of American Administrative Law 88 Harv.L Rev
1667 (1975); R. Stewart, The Development of Administrative and Quasi-Constitutional Law in Judicial
Review of Environmental Decision Making; Lessons From the Clean Air Act 62 Iowa L Rev 713 (1977);
A. Aman, Administrative Law in a Global Era: Progress, Deregulatory Change & The Rise of the
Administrative Presidency 73 Corn L Rev 1101 (1988).
26Motor Vehicle Manufacturers Assn.vState Farm Mutual Automobile Insurance Co.463 U.S. 29, at 42-3
(1983).
27Paul Craig,EU Administrative Law (Oxford University Press, 2006), Chap. 13.
28See, eg, Case 57/72, Westzucker GmbH v.Einfuhr-und Vorratsstelle fr Zucker [1973] ECR 321; Case
98/78, Firma A. Racke v.Hauptzollamt Mainz [1979] ECR 69.
29 Case T-13/99, Pfizer Animal Health SA v. Council [2002] ECR II-3303. See also Case T-70/99,
Alpharma Inc v. Council [2002] ECR II-3495.
30Ibid. para. 311.
31Ibid. para. 66.
32Ibid. para. 168.
33
-
8/10/2019 Paul_craig Law Fact and Discretion
34/34
33Case 138/79,Roquette Freres v. Council [1980] ECR 3333, para. 25; Cases 197, 200, 243, 245, 247/80,
Ludwigshafener Walzmuhle Erling KG v. Council and Commission [1981] ECR 3211, para. 37; Case C-
27/95, Woodcock District Council v. Bakers of Nailsea [1997] ECR I-1847, para. 32; Case C-4/96,
Northern Ireland Fish Producers Association (NIFPO) and Northern Ireland Fishermens Federation v.
Department of Agriculture for Northern Ireland [1998] ECR I-681, paras. 41-2.
34Case T-13/99, Pfizer, n. 29, para. 169.
35 Case T-342/99, Airtours plc v. Commission [2002] ECR II-2585; Case T-5/02, Tetra Laval BV v.
Commission [2002] ECR II-4381; Case C-12/03 P, Commission v. Tetra Laval.
36Craig,EU Administrative Law,n 27, Chap. 13.
37Craig,EU Administrative Law,n 27, Chap. 17-18.
38Craig,EU Administrative Law,n 27, Chap. 17.
39Shapiro, n. 23, 33-6.
40Stewart, The Reformation of American Administrative Law, n. 25.
41Case T-7/89, Hercules v. Commission [1991] ECR II-867, I.B.1. See also, Case C-344/98, Masterfoods
Ltd v. HB Ice Cream Ltd [2000] ECR I-11369, AG Cosmas, at para. 54.